More Tales of Pirx the Pilot (26 page)

“The new brass would like you involved in the work of the committee.”

Romani looked less drained than the day before, more upbeat. Simple logic told Pirx that even the mutually antagonistic “Martians” of Agathodaemon and Syrtis would close ranks if the “new brass” tried to railroad the proceedings.

The new committee had eleven members. Hoyster stayed on as chairman, if only because the committee couldn’t be chaired on Earth. A board of inquiry whose members were separated by eighty million kilometers was a risky venture; the authorities’ agreement to undertake it could only have been made under pressure. The disaster had revived a controversy of political dimensions, one in which the project had been embroiled from the start.

They began with a general recap for the benefit of the Earthlings. Among the latter, Pirx knew only the shipyard director, Van der Voyt. The color screen, for all its fidelity, lent his features a certain monumentally—the bust of a colossus, with a face both flaccid and bloated, full of imperious energy and shrouded by smoke rings from an invisible cigar (his hands were off screen) like some burnt offering. Anything said in the hall reached him after a four-minute delay, followed by another four-minute interval for the reply. Pirx took an immediate dislike to the man, or, rather, to his pompous presence, as if the other experts, whose faces flickered occasionally on the other monitors, were merely dummies.

After Hoyster came an eight-minute interval, but Earth momentarily demurred. Van der Voyt asked to see the
Ariel’s
tapes, a set of which lay by Hoyster’s microphone. By now, each member of the committee had received a dubbed set. Not that they were much help, since the tapes covered only the last five minutes of the landing sequence. While the camera crew relayed Earth’s set, Pirx fussed with his own, skipping the tapes with which, thanks to Haroun, he was already familiar.

The computer had reversed the landing procedure in the 339th second, shifting not to ordinary lift-off, but to an escape maneuver, as if in response to a meteorite alert, though it looked more like frantic improvisation. Whatever the sequence of events, Pirx attached little importance to the wild curve jumps on the tapes, which proved only that the computer had gagged on its own concoction. Of far greater relevance than a post-mortem of the ship’s macabre end was the cause of a decision that, in retrospect, was synonymous with suicide.

From the 170th second onward, the computer had functioned under enormous stress, showing signs of extreme informational overload, a piece of wisdom gained easily in hindsight, now that the final results were in. Not until the 201st second of the maneuver had the computer relayed the overload to the cockpit—to the human crew of the
Ariel.
By then, the computer was glutted with data—and kept demanding more. The tapes, in short, raised more questions than they solved. Hoyster allowed a ten-minute break for perusal of the tapes, then opened up the floor to questions. Pirx raised his hand, classroom-style. But before he could open his mouth, Engineer Stotik, the shipyard supervisor in charge of offloading the hundred-thousand-tonners, said that Earth should take the floor first. Hoyster wavered. It was a nasty ploy, beautifully timed. Romani asked for a point of order, declaring that if they were going to disrupt the proceedings by insisting on equal rights, then neither he nor anyone else from Agathodaemon planned to stay on the committee. Stotik yielded the floor to Pirx.

“The model in question is an updated version of the AIBM 09,” he began. “I’ve logged about a thousand hours with the AIBM 09, so I can speak from experience. I’m not up on the theory, only on what I’ve needed to know. We’re dealing here with a real-time data processor. This newer model, I’ve heard, has a thirty-six percent larger memory than the AIBM 09. That’s quite a bit. On the evidence, here’s what I think happened. The computer guided the ship into a normal landing sequence, then started overloading, demanding from the sub-routines more and more data per time unit. Like a company commander who keeps turning his combat soldiers into couriers: by the battle’s end, he might be extremely well informed, but he won’t have any soldiers.

“The computer wasn’t glutted; it glutted itself. It overloaded through the escalation—it
had
to, even with ten times the storage capacity. In mathematical terms, it reduced its capacity exponentially, as a result of which the ‘cerebellum’—the narrower channel—was the first to malfunction. Delays were registered by the ‘cerebellum,’ then jumped to the computer. As it entered a state of input overload, when it ceased to be a real-time machine, the computer jammed and had to make a critical decision. It decided to abort the landing; that is, it interpreted the interference as a sign of imminent disaster.”

“A meteorite alert, then. How do you explain that?” asked Seyn.

“How it switched from a primary to a secondary procedure, I don’t know. I’m not sufficiently at home with the computer’s circuitry to say. Why a meteorite alert? Search me. But this much I do know:
it
was to blame.”

Now it was Earth’s turn. Pirx was sure Van der Voyt would attack him, and he was right. The flabby, fleshy face, simultaneously distant and close up, viewed him through the cigar haze. Van der Voyt spoke in a polite bass, his eyes smiling, benignly, with the all-knowing indulgence of a professor addressing a promising student.

“So, Commander Pirx rules out sabotage, does he? But on what grounds? What do you mean,
‘it
was to blame’? Who is
it?
The computer? But the computer, as Commander Pirx said himself, remained fully functional. The software? But this is the very same program that has seen Commander Pirx through hundreds of landings. Do you suspect someone of having monkeyed with the program?”

“I’ll withhold comment on the sabotage theory,” said Pirx. “It doesn’t interest me right now. If the computer and the software had worked, the
Ariel
would still be in one piece, and we wouldn’t be having this conversation. What I’m saying is that, going by the tapes, the computer was executing the proper procedure, but in the manner of a perfectionist. It kept demanding, at a faster and faster clip, input on the reactor’s status, ignoring both its own limitations and the capacity of the output channels. Why it did this, I couldn’t say. But that’s what it did. I have nothing more to add.”

Not a word came from the “Martians.” Pirx, poker-faced, registered the gleam of satisfaction in Seyn’s eye and the mute contentment with which Romani straightened himself in his chair. After an eight-minute interval, Van der Voyt’s voice came on. This time his remarks were addressed neither to Pirx nor to the committee. He was eloquence personified. He traced the life history of every computer—from the assembly line to the cockpit. Its systems, he said, were the combined product of eight different companies, based in Japan, France, and the U.S. Still unequipped with a memory, still unprogrammed, as “ignorant” as newborn babies, computers traveled to Boston, where, at Syntronics Corp., they underwent programming. Each computer was then immersed in a “curriculum,” divided evenly between “experiments” and “exams.” This was the so-called General Fitness Test, followed by the “specialization phase,” when the computer evolved from a calculator to a guidance system of the type deployed by the
Ariel.
Last of all came the “debugging phase,” when it was hooked up to a simulator capable of imitating an infinite number of in-flight emergencies: mechanical breakdowns, systems malfunctions, emergency flight maneuvers, thrust deficiencies, near collisions… Each of these crisis situations was simulated in myriad variations—some with a full load, some without; some in deep space, some during reentry—increasing in complexity and eventually culminating in the most difficult of all: safely programming a ship’s course through a multibodied gravitational field.

The simulator, itself a computer, also played the role of “examiner,” and a perfidious one at that, subjecting the already programmed “pupil” to further endurance and efficiency tests, so that, although in actuality the electronic navigator had never piloted a ship, by the time it was finally installed aboard ship, it was more experienced, more flightworthy than the sum total of professional navigators. That is, the problems simulated during the bench-testing phase were too complex ever to occur in reality. And just to safeguard against the slightest imperfection, the pilot-simulator’s performance was monitored by a human, an experienced programmer with years of flight training behind him; Syntronics didn’t bother with pilots, only with astronauts at the rank of navigator or better, only with those, in other words, who had already logged a minimum of a thousand hours. With them rested the final decision as to which test, out of an inexhaustible range, the computer would be made to undergo next. The systems analyst specified the testing level and, by manipulating the simulator, further complicated the “exams” by simulating sudden—and potentially lethal—power blowouts, flare-outs, collision alerts, skin ruptures, communication breakdowns with ground telemetry … until the minimum standard of a hundred bench-testing hours had been achieved. Any model showing the slightest fallibility was sent back to the shop, like a flunked student who has to repeat a year.

Having in effect placed the shipyards beyond reproach, Van der Voyt, probably to counteract the impression of partiality, made an eloquent plea for an impartial inquiry. Next, a team of experts from Earth took the floor, thereupon unleashing a torrent of scientific parlance, flow charts, block diagrams, formulas, models, and statistical comparisons; and Pirx was chagrined to see that they were well on their way to turning the whole affair into an abstruse theoretical case study. The senior computer scientist was followed by Schmidt, the project’s systems engineer. Pirx very soon switched off, not bothering to stay alert for another round with Van der Voyt, which seemed progressively less likely. Not one reference was made to his own pronouncements, as if he had committed a
faux pas
—and the sooner it was forgotten, the better. By now they had reached the highest pinnacle of navigational theory. Pirx did not suspect them of malice; prudence dictated that they stay close to home. Throughout, Van der Voyt sat and listened indulgently. The strategy had worked: Earth was dominating the hearing. The “Martians” had been reduced to passive spectators; they had no surprise revelations up their sleeves. The
Ariel’s
computer was now electronic scrap, totally worthless as a source of clues. The tapes may have conveyed the
what,
more or less, but not the
why.
Not everything inside a computer can be monitored: one would need another, more powerful, computer, which, to be made foolproof, would again have to be monitored by yet another, and so on
ad infinitum.

Thus were they cast adrift on a sea of abstraction. The profundity of the disquisitions only obscured the fact that the tragedy extended far beyond the shipwreck of the
Ariel.
Automatic sequencers had been around for so long that they had become the basis, indeed, the inviolable premise, of all landing operations, and now this was on the verge of being snatched away. If none of the simpler, less foolproof computers had ever malfunctioned, why should a perfected, more sophisticated one fail? If that was possible, anything was possible. If the computer’s fallibility was open to doubt, there was no stopping the erosion of faith; then everything became mired in skepticism.

Meanwhile, the
Ares
and
Anabis
were Mars-bound. Pirx felt alone, on the brink of despair. The inquest into
the Ariel
had given way to a classic argument between theoreticians, and it was leading them further and further afield. As he looked up at Van der Voyt’s bloated, overblown face benevolently patronizing the committee, Pirx was suddenly struck by its resemblance to Churchill’s: the same look of apparent distraction, belied by a slight twitching of the mouth, betraying an inner smile provoked by a thought lurking behind heavy eyelids. What yesterday seemed unthinkable was now a foregone conclusion: a move to shift the burden of blame onto some higher force, onto the unknown, perhaps, or onto some theoretical omission, one that would require more extensive, long-term research. Pirx knew of similar, though less sensational cases, and knew the sort of passions such a disaster could arouse. Intensive efforts were already under way to reach a face-saving compromise, especially since the project, with its very existence now at stake, was ready to make concessions in exchange for support of the sort the shipyards could provide, if only by supplying a fleet of smaller cargo ships on favorable terms. In view of the high stakes—namely, the project’s survival—the
Ariel
calamity was an obstacle to be removed, if it couldn’t be immediately solved. After all, bigger scandals had been hushed up. But Pirx had one trump card. Earth had consented to his being on the committee because, as a veteran pilot, he was more at home with astronautical crews than anyone else present. He had no illusions: they were moved neither by his reputation nor by his credentials. Quite simply, the committee had need of an astronaut, an active one, a professional, all the better if that astronaut had just stepped ashore.

Van der Voyt smoked his cigar in a silence that, because it was dictated by prudence, lent him an air of omniscience. He might have preferred someone else to Pirx, but they had no excuse to bump him. If they now brought in an inconclusive verdict, and Pirx were to cast a dissenting vote, the result would be a lot of bad publicity: the press, having a nose for scandals, would pounce on it. The Riots’ Union and the Truckers’ Club wielded little power, but pilots made credible witnesses—they were, after all, people who put their lives on the line. So Pirx wasn’t at all surprised to hear, during the break, that Van der Voyt wanted a word with him. This friend of powerful politicians led off with a joke, calling their meeting a summit conference—at the summit of two planets. Pirx at times was subject to impulses that took even him by surprise. While Van der Voyt puffed on his cigar and lubricated his throat with beer, Pirx ordered a few sandwiches from the snack bar. What better way to get on an equal footing than by munching on a snack as he listened to the shipyard director in the communications room.